Author: Peter Fendrich
Date: 07:07:11 08/19/04
Go up one level in this thread
On August 18, 2004 at 23:23:14, Andrew Wagner wrote: >On August 18, 2004 at 18:53:23, Uri Blass wrote: > >>On August 18, 2004 at 16:59:32, Jonas Bylund wrote: >> >>>Let's say that someone were to cramp as much knowledge in to his/her program >>>with the sole purpose of making it stronger for _really_ long analysis/play, >>>thus not caring for the loss of speed, would this actually make the program >>>stronger for _really_ long games/analysis? >> >>It may make the program weaker because the program may have a lot of new bugs >>thanks for the new knowledge. >> >>> >>>I have a feeling that most programs are tuned and optimized for standard, rapid >>>and blitz play, not for 1 month games :) >>> >>>My point is that if there is indeed an increase in strenght if you to some >>>reasonable extend discard the speed vs. knowledge aspect, couldn't someone make >>>a long analysis version of their engine along with their normal engine? >> >>I do not think that the problem is a problem of speed. >>The main problem is that you think that programmmers know to give their programs >>productive knowledge and the only problem is that their program is going to >>become slower if they implement it. >> >>This is not the only problem and in a lot of cases the main problem is to know >>if some knowledge is productive and to implement things without bugs. >> >>Programmers have enough problems to find if a new version is better in standard >>games and if they try to do their best for that purpose they have not time for >>developing a special version that is better for long analysis. >> >>Uri > >I don't understand this response at all. The same could be argued about any new >feature to a chess engine -- "Null-move pruning is a bad idea because it could >introduce bugs and is not clear how to best implement it." The point is that >because of diminishing returns, the number reached by a knowledge-heavy engine >will probably not differ from that of a more classical engine if they both >search for a month, right? So I would have to give the advantage to the >knowledge-based engine, but that's just me. Uri's point is absolutely valid. The more sofisticated knowledge you add to your engine the harder it is to test. When you, above that, add the condition "1 month games" it is practically impossible to verify if the new piece of code it's ok. Numerous times I have added things to my engine, convinced that it was an improvement, just to find out that it did worse. It could be due to a bug or a mental black out that made me misunderstand how the code is changing the search tree. Whatever it is I have to catch it with sometimes extensive testing. I know that all engine programmers have done the same, maybe that I do it far too often however... :-) /Peter
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.